beta2 - pve-kvm: not correctly installed

D

donty

Member
Mar 31, 2009
42
0
6
I have a couple of issues here which feel like they should work out.

The server running proxmox stopped responding and required a reboot.

When I start a previously happy KVM machine instance it fails to start, running it from the command line shows:

vp4:~# qm start 101
VM is locked (backup)

I have rebooted the Proxmox machine and the vm is still not starting. When I run
pveversion -v
I get:
pve-manager: 1.4-6 (pve-manager/1.4/4340)
qemu-server: 1.1-4
pve-kernel: 2.6.24-15
pve-kvm: not correctly installed
pve-firmware: 1
vncterm: 0.9-2
vzctl: 3.0.23-1pve3
vzdump: 1.2-2
vzprocps: 2.0.11-1dso2
vzquota: 3.0.11-1

when I look at installing the pve-kvm I get

Package pve-kvm is not available, but is referred to by another package.
This may mean that the package is missing, has been obsoleted, or
is only available from another source
However the following packages replace it:
pve-qemu-kvm
E: Package pve-kvm has no installation candidate

Does anyone know if this is this the reason the VM is appearing locked or is there some way to unlock it?

Thanks!
 
donty said:
I have a couple of issues here which feel like they should work out.

The server running proxmox stopped responding and required a reboot.

When I start a previously happy KVM machine instance it fails to start, running it from the command line shows:

vp4:~# qm start 101
VM is locked (backup)

I have rebooted the Proxmox machine and the vm is still not starting. When I run
pveversion -v
I get:
pve-manager: 1.4-6 (pve-manager/1.4/4340)
qemu-server: 1.1-4
pve-kernel: 2.6.24-15
pve-kvm: not correctly installed
pve-firmware: 1
vncterm: 0.9-2
vzctl: 3.0.23-1pve3
vzdump: 1.2-2
vzprocps: 2.0.11-1dso2
vzquota: 3.0.11-1

when I look at installing the pve-kvm I get

Package pve-kvm is not available, but is referred to by another package.
This may mean that the package is missing, has been obsoleted, or
is only available from another source
However the following packages replace it:
pve-qemu-kvm
E: Package pve-kvm has no installation candidate

Does anyone know if this is this the reason the VM is appearing locked or is there some way to unlock it?

Thanks!
Click to expand...
see http://www.proxmox.com/forum/showpost.php?p=12008&postcount=1

run apt-get update and apt-get dist-upgrade.
try starting on the web interface and post the logs (from the web interface) why it is not starting.
 
OK, thanks that clears one thing up ;-) I didn't find it with the error as search terms.

Here's what shows around the time of a start attempt in the logs on the web interface, though they don't seem to help me much:

Oct 6 13:30:01 cron 4124 (root) CMD (test -x /usr/lib/atsar/atsa1 && /usr/lib/atsar/atsa1)
Oct 6 13:30:01 cron 4128 (root) CMD (/usr/share/vzctl/scripts/vpsreboot)
Oct 6 13:30:01 cron 4131 (root) CMD (/usr/share/vzctl/scripts/vpsnetclean)
Oct 6 13:35:01 cron 4155 (root) CMD (/usr/share/vzctl/scripts/vpsreboot)
Oct 6 13:35:01 cron 4157 (root) CMD (/usr/share/vzctl/scripts/vpsnetclean)
Oct 6 13:35:26 proxwww 4175 Starting new child 4175
Oct 6 13:40:01 cron 4180 (root) CMD (test -x /usr/lib/atsar/atsa1 && /usr/lib/atsar/atsa1)
Oct 6 13:40:01 cron 4183 (root) CMD (/usr/share/vzctl/scripts/vpsreboot)
Oct 6 13:40:01 cron 4186 (root) CMD (/usr/share/vzctl/scripts/vpsnetclean)
Oct 6 13:44:40 proxwww 3976 update ticket
Oct 6 13:45:01 cron 4222 (root) CMD (/usr/share/vzctl/scripts/vpsreboot)
Oct 6 13:45:01 cron 4224 (root) CMD (/usr/share/vzctl/scripts/vpsnetclean)
Oct 6 13:45:16 pvedaemon 4244 starting vnc console UPID:4244-250868:1254833116:vncview:0:101:root:5900:+XXXXXXXXXXXXXXXXXXXXXX
Oct 6 13:45:16 pvedaemon 4244 CMD: /usr/bin/vncterm -rfbport 5900 -passwdfile rm:/tmp/.vncpwfile.4244 -timeout 1 -c /usr/sbin/qm monitor 101
Oct 6 13:45:21 pvedaemon 4247 starting VM 101 on node 0 (localhost)
Oct 6 13:45:21 pvedaemon 4247 VM 101 start failed -
Oct 6 13:45:21 proxwww 4251 Starting new child 4251
Oct 6 13:45:27 proxwww 4253 Starting new child 4253
Oct 6 13:45:34 proxwww 4259 Starting new child 4259
Oct 6 13:46:25 proxwww 4297 Starting new child 4297

PS: I can create a new VM so that would seem to indicate a prob with the VM itself I presume?
 
Last edited:
I meant the log on the web interface, see "Administration/Logs/Tasks" - nothing here?
 
Nothing new there,

/usr/sbin/qm start 101
VM is locked (backup)
VM 101 start failed -

Same as the command line.

I tried to change the boot order to see if it would allow me to boot the w2k3 install and attempt to repair anything but when I tried to save it gives:

Error: unable to apply VM settings, command failed:
/usr/sbin/qm set 101 --localtime --freeze 0 --acpi 1 --kvm 1 --startdate --bootdisk ide0 --boot dc --vga --cpuunits 1000

Is there a way to 'release' the lock?
 
If you are sure that the backup already terminated, you can use

# qm unlock <vmid>

to release the lock (see man qm).
 
Yep, thank you, that did it.

I will now read the qm manual more closely.

My thoughts ran from the pve-kvm not installed error possibly causing the error 'vm locked (backup)' as I posted to begin with.

Once that was shown to be irrelevant then I suspected it was likely to be the vm and a need to unlock the vm. As the backup process itself was not showing as running, I wasnt sure what had actually locked the vm for the backup, hence the follow-up question.

So thanks for the help, now I need to discover why the underlying system stopped responding!

Cheers!
 
dietmar said:
If you are sure that the backup already terminated, you can use

# qm unlock <vmid>

to release the lock (see man qm).
Click to expand...


Cool.. I just bumped into this problem as well.. this solved it.. Thanks.
 
I had this exact problem. But after unlocking the VMs, they still won't start, and there don't seem to be any errors ... they just don't boot. The memory bar shows them only using about 10 Mb of RAM. Rebooting the whole PVE host doesn't seem to help.

I know the drill: look through error logs to see what is going wrong. I'm not even sure which logs can help me in this case.

I do have a backup host and snapshots of my VMs, but I want to have a go at fixing this before I bring things up on the backup.
 
Yeah, I have a Linux VM (102) and a Windows 2008 VM (101), and they both SAY they have started. The system log as shown in the PVM web interface shows "VM 102 started", (likewise VM 101), and no errors that I can see. But both VMs are using almost no RAM and no CPU time at all, and they certainly don't boot or run.

I notice in the "Virtual Machine Configuration" page that there is a tab for "Monitor" that gets one into a "qm" prompt. I don't know what, if anything, this console can tell me about what is going on.

At this point, I think I may have to nip over to my backup host, where snapshots of these 2 VMs are stored, and see if they start. Of course if it was last night's snapshot that bombed, that VM may not be usable, and doggone it, I haven't been keeping multiple snapshots.
 
I do have an earlier snapshot, thankfully, but here's another question: what would have precipitated this? Last night's snapshot, which is done to an NFS share on our backup host, did fail on VM 101, but this morning, BOTH VMs were stopped. 101 would not even try to start until I did the qm unlock thing, but 102 was already doing the "I'll startup, but not really" thing. I'm trying to think what the original screwup was. (Hosting system is on a nice new UPS too, so no power hiccup.)
 
....aaaaaand it gets worse. Using a qmrestore to bring up my VM101 Win2k8 on my backup hardware PVE machine, it looks just fine in the web interface, every part of every configuration area looks identical to the VM on the main hardware. Looks like the qmrestore has indeed duplicated the VM on the main machine. And, it looks like VM101 tries to start and boot. But it never finishes. I look in the VNC console window and the VM is asking me to hit F12 to choose a boot device (which I can't do) . So for some reason it can't see the boot drive.

Right now my whole rationale for building a VM hosting system is crumbling before my eyes.
 
You must log in or register to reply here.
Top Bottom